Create Searchable Documents Across a Mixed Matter Folder
By Schema LawPublished Updated
To create searchable documents across a matter folder, sort the files by whether they already carry text. Word files, spreadsheets, email bodies and exported PDFs do. Scans, photos and image attachments do not, and only those need text recognition. Convert that group, save each result beside its source, and test it with a search.
Which documents in a matter folder are searchable already?
Any file that stores characters is searchable already: word-processor documents, spreadsheets, presentations, plain text, the body of an email, and a PDF that was exported from one of those programs. These need no conversion. What they need is a search tool that is set to read them.
The test is the same for every type. Open the file, press Ctrl+F (Command+F on a Mac) and search for a word you can see. A hit means the characters are there.
On each converted scan, search the party names, a docket number, and the label of the first exhibit. Open the hit and compare it with the page image. In a records set, search one medical provider as well. A subpoena you can read on the picture and cannot find from the folder goes back on the list to convert again.
| File type | Carries text? | The catch to check |
|---|---|---|
| Word, RTF, plain text | Yes | Pictures pasted into the document are not text |
| Excel, CSV | Yes | A screenshot of a table pasted onto a sheet is a picture |
| PowerPoint | Yes, in text boxes | Slides built from screenshots hold pictures |
| Email body | Yes | Attachments are separate files with their own answer |
| PDF exported from an office program | Yes | Pages inserted from a scanner later are pictures |
| PDF from a scanner, copier or fax | Only if a text layer was added | Test a middle page, not the first |
Whether a search across the folder finds those characters is a second question, and it belongs to the search tool. Microsoft's search indexing FAQ for Windows says "For files with text, their contents are indexed to allow you to search for words within the files", and that each file type can be set to index "properties only, or properties and content". The same page gives the path to the File Types tab where you can see the setting for each extension. If a folder search misses a word you can find inside a file, look there before you convert anything. The post on making PDFs text searchable for a folder search covers what an index reads from a PDF.
File types that hide their words from search: scans, pictures and image attachments
Three groups hide their words: PDFs made of page pictures, image files such as JPEG, PNG and TIFF, and pictures that sit inside or travel with another file. Each shows you text on screen and stores none, so a search passes over it without a message.
Silence is the problem. A search that finds twelve hits looks complete. Nothing tells you that forty more files were never read.
Scanned PDFs
A PDF from a scanner, a copier or a fax line is a stack of pictures unless something added a text layer. The post comparing native, scanned and searchable PDFs explains the three kinds. In a matter folder, look hardest at files with names a device gave them, such as SCAN_0042.pdf, and at anything the other side sent.
Picture files
Photos a client took of a notice, a screenshot of a text message thread, a TIFF from a records office. The file name is searchable and the content is not.
Pictures inside other files
These are the ones a sort by file type misses.
- A Word document whose signature page was pasted in as an image.
- A PDF assembled from a typed letter and a scanned enclosure.
- An email with a photo placed in the body instead of attached.
- A spreadsheet with a screenshot of a bank statement on one sheet.
- A compressed archive such as a ZIP file, which hides everything in it until it is unpacked.
Sorting the folder
Make a listing of the folder with one row per file, and add two columns: "has text" and "action". Sort by extension. Image extensions go straight to the convert pile. Office extensions go to the leave pile, with a note to look for pasted pictures in the ones that matter. PDFs need testing one by one or in bulk; the triage table in the post on sorting a folder of PDF documents by whether they need OCR covers that step.
Creating searchable documents from each unsearchable file type
Each unsearchable type has its own short route, and they all end the same way: a file that holds recognized text, saved beside the source and tested with a search. The table gives the route for each.
| What you have | Route to a searchable document | Save as |
|---|---|---|
| Scanned PDF | Run text recognition on the PDF | Searchable PDF |
| Single JPEG or PNG | Run recognition on the image | Searchable PDF, or a text file beside the image |
| Many loose page images of one document | Join them in page order into one PDF, then recognize | One searchable PDF |
| TIFF, one page or many | Convert to PDF, then recognize | Searchable PDF |
| Part-typed, part-scanned PDF | Recognize the picture pages and leave the typed pages as they are | Searchable PDF |
| Word file with a pasted picture page | Export to PDF, then recognize the picture page | Searchable PDF, with the Word file kept |
| Photo placed in an email body | Save the picture out as a file, then treat as an image | Searchable PDF filed with the email |
| ZIP archive | Unpack to a subfolder and sort the contents as a small folder | Depends on what is inside |
Three of those rows have their own guides. Picture files are covered in the post on turning a scanned image into text, which also says what to do with types a tool refuses. TIFF has its own procedure in the post on converting TIFF to searchable PDF. For any PDF, the hub post on creating a searchable PDF helps you choose a route by what you start with.
Convert a PDF into a searchable format, step by step
To convert one scanned PDF into a searchable format, copy the scan, run recognition on the copy with the tool your firm has approved for the matter, and save the output as a searchable PDF beside the source. The guide to making a scanned PDF searchable sets out that procedure step by step. In a matter folder, two things are added to it.
- Tie the pair together by name. Give the searchable copy the source's name with
_ocradded, so the two sort side by side and nobody mistakes one for the other. - Close the row. Compare the page count with the source, search the new file for a word from a middle page and a word from the last page, then mark the file's row in your folder listing as done, with the date.
In Schema Law, that is: sign in, upload the PDF, PNG or JPEG (up to 20 MB), look through the parsed pages, and export. Searchable PDF is an export format on every plan, and the Free plan's one document a day of up to three pages is enough to try the route on a short scan. The how it works section shows the steps and the pricing page lists the page limits for longer files.
A whole folder of scans is a different size of job. The post on bulk OCR for PDF folders covers running many files while keeping names and subfolders intact.
Should everything become a searchable PDF, or stay in its own format?
Leave files that already carry text in their own format, and make searchable PDFs only from the files that are pictures. Converting a Word file or a spreadsheet to PDF adds nothing to search and takes away the things the original format was for.
What a blanket "PDF everything" rule costs:
- Spreadsheets lose formulas, hidden columns and the ability to sort. A wide sheet printed to PDF is cut across pages.
- Word files lose tracked changes and comments unless they are deliberately printed.
- Emails can lose header details and the link to their attachments, depending on how the conversion is done.
- The folder doubles in size, and every document now exists in two forms that can drift apart.
There are good reasons to convert a particular file anyway.
| Situation | Keep the original format | Also make a PDF |
|---|---|---|
| A draft still being edited | Yes | No |
| A final version sent to the other side | Yes | Yes, as the record of what was sent |
| A document that will be an exhibit or a filing | Yes | Yes, in the form the court requires |
| A spreadsheet the expert will work in | Yes | No |
| A file in a format few people can open | Yes | Yes, so colleagues can read it |
| A scan, a photo, a fax | Yes, as the original image | Yes, a searchable PDF |
The rule underneath the table: a searchable PDF is the right form for a picture of a page, because it keeps the picture and adds the words. For a file born with characters in it, a PDF is a copy made for a purpose, and you should be able to name the purpose.
What a court wants in a filed PDF is a separate matter with its own rules; the post on searchable PDFs for court filing covers those and tells you to check your own court.
One more choice sits inside this one. A scan can become a searchable PDF, which keeps the page picture, or an editable file such as Word, which keeps only the words and layout. For a matter folder whose purpose is finding and proving things, the searchable PDF is the record copy. Make the Word version when someone needs to reuse the language, and file it as a working copy.
Emails with scanned attachments: making the attachment searchable too
Treat the attachment as its own document. A search of the mailbox can read the message and, depending on the mail program, some attachments, but a scanned PDF or a photo attached to an email has no text for it to read until you add some.
Microsoft's guide to searching in Outlook says that "Outlook scans both email messages and many types of attachments for that word or phrase." That sentence does not say which types, and no mail search can find words that are not stored in the file. So test yours: pick an email whose attachment is a scan, search the mailbox for a word that appears only on a page of that attachment, and see whether the message comes back.
Finding the attachments that need work
The same Outlook page lists the search term hasattachment:yes for "Items that have attachments". Combine it with the client's or matter's name to list candidate messages. Other mail programs have their own filters; check their help.
A routine for one email
Take an email from a client, Anselm Druitt, with three attachments: a Word letter, a scanned PDF of a signed lease and a phone photo of a meter reading.
- Save the email itself to the matter folder in the format your firm uses for filed email.
- Save all three attachments to the same folder, with names that tie them to the message: the date, the sender and a short description.
- Leave the Word letter alone. It has text.
- Run recognition on the scanned lease and save a searchable PDF beside the scan.
- Run recognition on the photo, or note its content in the file name if it is a single reading.
- Record in the folder listing which message each file came from.
Step 6 is what keeps the family together. Once an attachment is saved out and converted, nothing in the searchable copy says which email delivered it, and that link can matter as much as the content. The post on document unitization explains parent documents and attachments in a scanned set.
Do not replace the attachment inside the stored email with your converted copy. The message as received is the record; the searchable copy is a tool you made from it.
How do you keep a mixed matter folder searchable as documents arrive?
Check every new file on the day it enters the folder, whatever route it came by. One question, "can I search inside this?", asked at the door is cheaper than a sweep of four hundred files before a deadline.
A five-line intake rule
- Test. Open the file and search for a word you can see.
- Convert if it fails. Scan, photo or image attachment: make a searchable copy now.
- Keep both. The file as received, and the searchable copy beside it.
- Name them as a pair. Same base name, so they sort together.
- Log it. One row in the folder listing: file, source, date, has text, converted by.
Where files come from
Files arrive by several channels, so tell whoever handles each one what to look for.
| Channel | What to check on arrival |
|---|---|
| Attachments saved out and tested, not only the message | |
| Client portal or file-sharing link | Downloads that are photos or scans; archives that need unpacking |
| Opposing counsel | Whether PDFs have text on every page; whether images came with separate text |
| The office scanner | Whether its profile adds a text layer, and whether that layer is any good |
| Phone photos from a client or investigator | Each picture is an image file with no text |
| Court or agency download | Test it like any other PDF; do not assume |
Paper that comes through the post has a routine of its own, covered in the post on scanning incoming mail to searchable PDF.
A monthly sweep
Intake rules slip. Once a month, or before any search you intend to rely on, sort the folder listing by date added and test the new files that have no entry in the "has text" column. The guide to checking whether a PDF is searchable on every page gives a method for a set.
Before you rely on a search
A matter folder is searchable only up to the last file somebody checked. Before a search whose result you will act on, confirm three things: the listing has no blank rows, the search tool is set to read the contents of every file type in the folder, and a search for a word you know sits inside a converted scan brings that file back. If the third test fails, the gap is in the search tool or in the file, and you need to know which before you trust a result of "nothing found".